Rachel Kit

Face Your Art Challenge

[Labrin @Labrin]

3年前写过一个 solver 自己接 LP 用

今天正式将它配合 0x sellter 在 Robinhood chain 上提供

我把它叫做 SHEEP CHOICE

1. 不虚报价格

2. 测试下，实际执行价格优于已测试过的聚合器

仓库：https://t.co/sXbq6aoH0H

网页版本直接使用： https://t.co/ZGIgSPUD1g

[TimSEO @Timseo_]

Arrêtez de dépenser votre argent dans ce genre de scams. Faites tout le contraire :

1. choisissez une niche ultra-spécifique.
2. récoltez le plus de données possible sur cette niche.
3. identifiez 3 à 5 mots-clés business.
4. demandez à Claude de vous lister 10 micro-intentions pour chaque mot-clé business.
5. créez le meilleur contenu possible autour de ces 30 contenus (3 × 10).
6. publiez une vidéo YouTube par mot-clé business.
7. montez une communauté autour de cette thématique.

voilà

[Hridoy Reh @hridoyreh]

10 AI SEO tools that can be used:

1. Ahrefs
2. SEMrush
3. Surfer SEO
4. Outrank
5. Jasper AI
6. Rankability
7. Byword AI
8. Alli AI
9. Profound
10. SparkToro

[Hridoy Reh @hridoyreh]

I got 12 quality backlinks for free.

The "SEO Wins" hack I used:

1. Find statistics pages in your niche.
2. Check them with a broken link checker.
3. Find the dead sources they reference.
4. Create an updated version on your site.
5. Email the site owner.

Only hard work will get you links...

[小奇说 @xingbugengming]

https://t.co/j4he7fGMCB

[Berryxia.AI @berryxia]

这个APP小黄鱼的骗子估计该恨死我了！
上周朋友线下收二手 Mac，差点亏大了。

卖家说得特别好听：无 ID、无暗病、序列号官网也能查。
结果回家一细测，发现机器还藏着企业 MDM 绑定。

这种机器后面转手很麻烦，朋友差点直接亏两千。

很多人买二手 Mac，其实验机逻辑都错了。
不是只看外观，也不是只查个序列号就完事。

我最近做了一个本地验机工具，叫 MacCheck 验机宝。
它直接在 Mac 本机运行，全程不上传数据。

打开之后，芯片、内存、电池循环次数、硬盘容量、安全芯片状态，都会自动扫出来。

键盘可以逐键检测，屏幕可以用纯色和灰阶查亮点坏点。
摄像头、麦克风、触控板、Touch ID、接口，也都能一项项核验。

更重要的是，像激活锁、Apple ID 登录状态、企业 MDM 绑定这种高风险问题，会直接高亮提醒。
不用你自己到处翻设置、查资料。

验完之后，还能一键导出 PDF 或长图报告。
以后复盘、对比、留证，都很方便。

以前买二手 Mac，很多时候靠卖家一张嘴。
现在至少可以自己把整机硬件、隐藏锁具、核心风险先查一遍。

如果你最近要买二手 Mac，或者做回收验机，
可以试试这个工具：MacCheck 验机宝。

我已经将项目免费开源到Github，记得一键三连和Star给我啊！

我使用Swift语言开发，仅需2.7M，立刻下载使用吧！

地址下面👇🏻：https://t.co/etIYJEyOSx

[ccjing | (✱,✱) 🐬 @ccjing_eth]

大多数用户现在用的 Codex（也就是ChatGPT客户端）是阉割版

被偷偷阉割了生图和webSerach功能

如何获得满血版Codex：

直接把下面这一串提示词发给Codex让他帮你加修改：

帮我修改配置文件，打开ChatGPT客户端的生图和Webserach功能

记得打开完全访问

[Rachel Kit @kit_rachel]

@OneLegchris "Then mine the comments. Scrape a few hundred users. What is your data source?"

[Chris - Local SEO @OneLegchris]

Most people fight to rank their homepage for the keyword. I dont. I build a dedicated service page and rank that. And before I build anything, I search the keyword, because just searching it hands you what is ranking and what it takes to beat them. Read the top profiles and services, write down what Google is bolding, because Google is bolding it for a reason.

Then mine the reviews. Crawl a couple hundred profiles and read the negative ones, because thats where the pain points live and how people actually search. The complaints become your FAQ. Roots left below grade, crews that dont clean up. And list fifteen services with a description on each: that alone beats close to ninety percent of your competition.

On the service page, keyword in the H1 with the city and state, bold it, use listicles, answer cost and process. Hide the SEO below the fold in scrollable review cards so the customer never feels it. Then link your areas and entities to their Knowledge Graph IDs, Wikipedia, and Wikidata, so Google pulls in everything it already knows and relates it to you for free.

[Rachel Kit @kit_rachel]

@kaikaibtc 底裤都被你扒光了

[K三 凯 @kaikaibtc]

一个开发者，揭开了一件离谱的事 - Meme发射台生意的暗线

最近在陪跑一个 Agent + 税币的发射台项目，做竞调的时候把市场上几个主流 launchpad 的税费机制都研究了一遍。

看完之后有个感受：发射台这个生意，表面卷的是流量、资产和玩法，长期看卷的是费用结构。

🤖pumpfun内盘 Bonding Curve 每笔交易总费率 1.25%，其中 0.95% 给 Protocol，0.30% 给 Creator；毕业后，交易费按市值/池子规模动态下降，费率逐步降到 0.30%，Protocol fee 就是平台收入的重要来源。

👿Fourmeme分了两种模式：普通 token 和 Tax Token。Tax Token 允许 设置交易税率，税收进入项目方地址，平台本身不从中抽成。https://t.co/9fr9Psy1Xt 的收入主要来自token 创建费、毕业费这类一次性收入，和交易税是两条线。

🟦Base 链上的几个发射台（Virtuals、Zora、banker）就有点乱了——有的收创建费、有的收交易手续费、有的抽毕业流动性，费率也不固定，没有统一标准。

这些平台模式不一样，但共同点是：协议费收入已经是发射台的重要收入来源，而且都是明码标价的。

然后就是 Flap🦋，这波税币发射台里跑得最久的，在 BSC 上发育了两年，存在感越来越强。

它的税币机制设计得挺巧妙：creator 可以给 token 自定义交易税率（1%、3%、5%、10%），交易产生的税进入一个 TaxSplitter 合约，再分配到营销、分红、回购、金库等用途。

我本来想重点学习一下它税币的分配机制，看看费用结构，结果发现了几个值得讨论的小细节。

1⃣细节1，税收并非官方说的100％

Flap 的税币机制大概是这样：创作者可以给 token 设置交易税，比如 1%、3%、5%、10%。用户交易时产生税费，这笔税再进入营销、分红、回购、金库等用途。

但flap税币的 TaxSplitter合约中，都出现了同一个参数：
feeReceiverAmount = balance − balance × 9000 / 10000

意思是，税收进入后会被拆成两部分：90% 进入项目方设置的 recipient，另外 10% 进入 feeReceiver。

而这个 feeReceiver，指向同一个地址：
0x8a08D98CBB218fceB318Ecf3aBc1BA43D8A7aB0E

也就是说，交易者看到的税收去向是Marketing/Burn/Dividend/Liquidity 四项，官方说法是100%，但协议费构成是隐形的。税收的10％被Flap抽走了，但却没有任何提示和信息高知。

2⃣细节2，两年吞了13,547个BNB

用Agent顺着 0x8a08查资金来往，发现从 2024年7月收到2026年7月，连续两年，累计进账13,547 个 BNB，大概总价值10M吧。

这13,547 BNB里，来自税币10%抽成的部分约 7,900 BNB，大概价值5M；剩下是内盘 1% 曲线费，这个Fee是flap之前公布过的。

1,224 万笔流入，平均每笔 0.0011 BNB，不到一块钱。但连续两年这样流，竟然也流出了500万美元，说明这不是偶发收入，而是一套长期存在的协议级收费入口......

而且这些资金大部分进入两个Gnosis Safe多签地址，其中一个地址持有 6,912 BNB，价值约 460 万美元，多签签名人中还包含了 Flap 部署钱包。

平台收费当然不是问题，但当一条资金管道长期运转，并累计到几千 BNB 规模时，它就必须成为用户明确知情的产品规则。

3⃣细节3，Robinhood 链上同样的剧本

在查完 BSC 上的合约后，我又去看了 Flap 在 Robinhood 链上部署的合约，结果一模一样。同样的 TaxSplitter 逻辑和feeReceiverAmount 参数，税费90:10 分账比例。

说明两件事：第一，这不是 BSC 合约的偶然疏漏，而是在多链部署时复用的同一套协议逻辑；第二，Robinhood 链上的 Flap 用户和 BSC 用户一样，对这个 10% 的存在并不知情。

4⃣细节4，流水的盘子，不对称的信息

平台协议收入进入团队多签，可以有很多合理解释，比如产品开发、运营维护、安全支出、市场推广。

但合理不等于无需解释，恰恰因为它可以被合理解释，所以更应该解释清楚。

Flap 官方在各个渠道的口径都说税费是100％，但合约写的是90％，如果是遗漏的话，希望官方能补全一下，以免误解。

普通创作者和 holder，如果不懂合约代码，根本不可能知道这个机制的存在，而他们恰恰是最信任 Flap 的那群人。

Web3 常说链上能看到一切，去中心化，代码即规则。

但对绝大多数用户来说，前端、官方解释、社区公告才是规则，作为一个面向 meme 散户的平台，不能只对专业用户透明，协议收入也不是问题，讲清楚写明白就OK。

可能发射台项目真正需要卷的，不只是功能和流量，还有规则透明度。用户并不要求平台永远完美，但至少希望在涉及收费、分账和资金流向时，项目方能把规则一次讲明白。

对 Web3 产品来说，真诚不是态度包装，而是一种产品能力：把钱讲清楚，把规则讲清楚，把用户本该知道的东西放到阳光下，这才是产品应有之道。

[季白羽 @vbjby3]

https://t.co/AJecssA9vW

[懂币猫 @dongbimao]

#止损

一说到止损，很多朋友总是很纠结，到底要怎么设置，有时候干脆不设置，核心来说止损的关键不在于你天天纠结多少个点

止损的关键在于：
你要如何果断的执行止损

进一步：
止损多少个点你不纠结，你要先问自己
1个点
2个点
3个点

只要做到：截断亏损就可以

止损重要的是逻辑正确，而不是寻求什么完美的点位，世界上并没有完美的止损点位，只要能让我们留在牌桌上的止损就是最好的止损

[daedalus @aiofmgod]

100,000,000 people speak German & the number of fake AI girls who speak it back is functionally zero

Every operator on the internet built the same exact girl... a 22-year-old American who types "hey babe" in English to broke men in Ohio

Meanwhile the richest lonely men on earth... Germany, Japan, Korea, the Gulf... have nobody speaking their language & nobody competing for their wallets

The tool that unlocks all of them costs $22/mo & most operators already pay for it

Every ban wave, every rising acquisition cost, every "the space is saturated" complaint on this app is one problem wearing different costumes... 10,000+ operators fishing in ONE pond: English-speaking men

The pond isn't overfished. Everyone is crowded in the same corner of it

& if you think foreign men won't pay for conversation, Japan built an entire physical industry on it. Hostess clubs... men pay hundreds of dollars a night for a woman to pour drinks & ask about their day. The law literally prohibits her from touching them. Conversation IS the product & Japanese men have paid premium for it since the 1960s

The play:

- Build the girl local, not translated. German name, Munich apartment backstory, complains about her train being late again. One local detail does more than a perfect body
- ElevenLabs v3 speaks 74 languages. Same $22/mo tool. Her voice notes come out in native German
- You don't need to speak a word of it. Every message runs through a translation layer before it hits your screen & before it leaves yours. No man alive clocks a 60-second delay from a girl who "just got out of the shower"
- Price HIGHER than you would in English. When a man has never seen a girl like her speak his language, she isn't a commodity... she's the only one

Why nobody runs this: every course, every thread, every Discord in this space teaches in English, so every student gets dumped into the same English market to fight over the same men

The operators who went to Reddit early stopped fighting for scraps. The ones who go foreign-language get an entire hemisphere to themselves

Interesting times

Fill out my form in bio or pinned tweet

If you're qualified I'll teach you this EXACT SYSTEM

[Bryan Ng @boomerrbryan]

A guy in Portugal is running the #1 retirement planning channel in Brazil and he doesn't speak Portuguese

He took every script from a proven English retirement channel, translated them with AI, had an AI character deliver them in native Brazilian Portuguese, and uploaded them into a market of 220 million people where zero serious channels existed

The English channel he copied from has 400K subscribers and a knife fight in the comments with 30 other channels covering the same topics. The Portuguese copy has 86,000 subscribers and basically no competition at all. Same scripts. Same topics. Same thumbnail formats. One market is a warzone, the other is an open field

Monthly revenue on the Portuguese channel: roughly $14,000. His production cost per video: under $10

Here's the play:

English YouTube is the most competitive content market on earth. Every niche has 50 channels. Every topic has been covered 200 times. Breaking through requires either massive production budgets or years of grinding

Spanish YouTube serves 500 million native speakers. German serves 100 million. Japanese serves 125 million. Portuguese serves 260 million. In most niches across all of these languages, the competition is 5% of what it is in English

A retirement video that did 2 million views in English addresses the exact same anxiety as a retirement video in Spanish. A 60 year old in Ohio worrying about running out of money has the same fear as a 60 year old in Madrid. The emotion is universal. The content translates 1 to 1. The competition doesn't

The system:

1) Find a proven English channel in a high-RPM niche. History, finance, health, geopolitics, retirement, true crime. Screenshot their top 20 videos by views

2) Search the same topics on YouTube in Spanish, Portuguese, German, or Japanese. Filter by the target language. Count how many serious channels exist. In most niches the answer is 1 or 2. Some niches have zero

3) Translate the scripts with Claude or DeepL. Not word for word. Localize the references. The English version mentions the IRS. The Spanish version mentions Hacienda. The English one references 401k. The Spanish one references Plan de Pensiones. The structure and insights stay the same. The specifics adapt

4) Generate an AI character who looks and sounds native to the target market. A 55 year old Spanish-speaking man for the retirement niche in Latin America. A calm Japanese woman for the health niche in Japan. The character needs to match the audience's demographic exactly

5) Clone the thumbnail format from the English original. Swap the text to the target language. The thumbnail design already proved it gets clicks. The only variable is the language

6) Post 3x per week into a market that has almost nobody serving it

The RPMs in non-English markets are lower ($5 to $15 vs $15 to $45 in English) but the competition is so thin that views come 10x easier. Lower RPM x 10x views = more total revenue than fighting for scraps in English

The English channels will never know you exist. They're not searching YouTube in Portuguese. They're not tracking their analytics by language. Their content is being repurposed in markets they've never considered and they'll never find out unless someone tells them

nobody is going to tell them

we currently have a waitlist for the tool that replaces your entire production team and makes YouTube actually work for your business. waitlist link is in bio. spots are limited.

[nofad @nofadsec]

https://t.co/1yEDKd7EpB

[Bober_smart @Bober_smart]

A 27-year-old guy from China bought 100 Mac minis and turned his apartment into a server hub

Instead of furniture, the space is filled with neat racks where a hundred Mac minis hum like a living organism, processing gigabytes of data in a continuous stream

Many small AI startups in Asia cannot afford to train models on NVIDIA H100 servers: it is insanely expensive

He provides his Mac minis for running and fine-tuning small, highly specialized models. He takes on the dirty work of data labeling and quality assurance by using his network of agents that check each other's work

> His initial investment was $59,900
> He was able to recoup his costs in 3 months

Right now, his average income is $25,000–$30,000 per month

He is not a genius: he simply found a niche with high demand and created an offer

[雨哥向前冲 @xiangxiang103]

刚发现一个神站！🚀

https://t.co/mJG5tiCV9S

里面全是高端AI网站提示词，复制粘贴给你的AI就能秒出带动画的超美落地页、SaaS、作品集……

几分钟做出以前要花几天做的效果，太香了！

[所長 @ZO000I]

老马又用Grok放福利了？白嫖一个美国号码？

现在几分钟就能搭建一个 AI 电话代理，还会提供一个美国号码用于接听电话

教程⬇️

① 打开：https://t.co/t8wBpDp4Yk

② 登录 xAI（Grok）账号

③ 点击 Create Agent

④ 写好 Prompt（决定 AI 怎么接电话）

⑤ 选择声音（支持 Voice Clone）

⑥ 点击 Deploy

完成后即可获得一个用于 Voice Agent 的美国号码。

你可以直接：

✅ 浏览器测试通话
✅ 手机拨打体验
✅ AI 自动接电话
✅ 自动回复、记录留言、处理预约

整个过程不用写代码，也不用自己搭服务器。

目前官方费用约 $0.06/分钟（语音 + 电话线路）。

以前搭一套 AI 电话系统，要买号码、接语音平台、写接口。

现在，一个 Grok轻轻松松解决

AI 电话代理，已经开始普及了@grok

[Indie Fox @indie_maker_fox]

真爽，现在我想要集成什么数据，简单配置下就能用了

从此 agent 能够看到我的各个应用数据并帮我执行操作

接下来需要搞些优秀的 skills 和 automation workflow

截图是 Echo，桌面端的 agent 产品，目前还在开发中